New England Patriots at Buffalo Bills. Lean: New England Patriots +7. Pick: Under 48.5.

Spread.Pass with a lean to Buffalo -6.5 at -120. Market first, the move from DraftKings -3 to -6.5 is best explained by Buffalo's 3-0 start and New England's offensive collapse, including Maye's turnovers, the loss of A.J. Brown and mounting line injuries. That move has already incorporated much of the public case, and the model merely matches the current consensus at Bills -6.8, so it supplies no independent spread edge. The remaining directional case is matchup specific: Buffalo's pass rush has been productive, while New England entered the week 28th in sacks allowed per play and again listed Morgan Moses as a nonparticipant. Buffalo also faces a depleted Patriots defense, but Christian Gonzalez's status and Buffalo's own injuries prevent a bet at a key number adjacent to seven. I would only retain the lean at the captured -6.5, not at -7.

Total.Pass with a lean to under 48.5 at -105. Market first, the total fell from the earlier DraftKings 49.5 and FanDuel 50.5 openings because New England has scored 36 points through three games, lost Brown to IR and brought major line concerns into Buffalo. That is a sensible correction, not a fresh inefficiency. The under case remains that Maye's recent mistakes, a weakened receiving group and Buffalo's front can leave New England unable to contribute enough to a high total. Buffalo's receiver injuries also trim some ceiling. The opposing case is substantial: Buffalo scored 24 despite five turnovers last week, scored freely in Weeks 1 and 2, and the current mild forecast offers no weather brake. Our 46.68 model number points under, but its documented holdout performance was worse than market, so this stays a price sensitive lean only.

Where our number looks wrong.The model's -6.8 agrees with the market, but its preseason based Patriots rating is likely too generous because it does not price A.J. Brown's IR absence, New England's current line attrition or Maye's turnover problems. That does not establish a playable Buffalo edge after the market's large move.

• Patriots announced Brown's IR placement and stated that Week 6 is the earliest possible return.
• Patriots analysis detailed the 35-6 Week 3 loss and Maye's two interceptions and lost fumble.
• Bills recap reported a 24-16 Week 3 win despite five Buffalo offensive turnovers.

Spread.Market first: the strongest case for Bills -7 is that New England has fallen apart since the lookahead number of Bills -3. Drake Maye has 1 touchdown and 6 interceptions with a league high seven turnovers, A.J. Brown is on injured reserve until Week 6, the right side of the offensive line is gone (Onwenu and Van Roten on IR, Moses did not practice), and Christian Gonzalez missed Wednesday. Buffalo is 3-0 and scoring 33.7 points a game. The lean to New England is about the size of the move, not a hidden edge: the line traveled roughly four points on one 35-6 blowout plus injuries, and it now sits on the key number 7. Buffalo's defense has allowed 31, 31 and 16, ranks 25th in opponent EPA per play per SI, and will likely play without CB Christian Benford (toe, walking boot, DNP). Division familiarity and a Patriots defense that allowed 8 points a game in Weeks 1 and 2 give the plus 7 some cushion. This is a small lean at plus 7, not a bet.

Total.Over case: Buffalo games have totaled 67, 72 and 40, both secondaries may miss their top corner (Benford, Gonzalez), the forecast is mild and dry, and Maye turnovers hand Buffalo short fields. The over is 7-3 in the last 10 meetings per SI. Under case: New England has scored 10, 20 and 6 with a broken right side of the line and no A.J. Brown, and its three games totaled 23, 23 and 41. Buffalo leans on James Cook at 6.0 yards per carry, which shortens games when it leads. Our model says 46.7, and after its known upward bias of about 1.2 points it sits near 45.5, below 48.5. The market has already dropped from 50.5 at FanDuel, so part of this is priced. I lean under 48.5 because I doubt the Patriots can supply their share of a high total, but not enough to bet.

Where our number looks wrong.Our number matches the market, so there is no gap to defend. The rating likely still carries too much of New England's 2025 offense (0.143 EPA per play with a healthy line and Maye at a 113.5 rating) and does not see the two IR guards, Brown's absence or Gonzalez, so if anything it may understate Buffalo; the reason to still lean Patriots is market overreaction, not the model.

• Benford in walking boot after Chargers game; A.J. Brown on IR
• Bills 25th in opponent EPA per play; Patriots ninth in net yards per play; BetMGM Bills -6.5 at time of writing
• Official Wednesday injury report for both teams
• Patriots 20-3 over Pittsburgh in Week 2
• Seahawks 13-10 over Patriots in Week 1
• Bills 41, Lions 31 in Week 2
• Bills 36, Texans 31 in Week 1
• Patriots defense allowed 8.0 points a game in Weeks 1 and 2 before the Jacksonville loss
• Over 7-3 in last 10 meetings; total 48.5

Spread.The strongest explanation for Bills minus 6.5 to 7 is simple: Buffalo is 3-0 at home with Josh Allen, New England is 1-2 and thin at receiver with A.J. Brown hurt and the right guard on injured reserve, and the Patriots were a 14-3 team last year so the market is not fully discounting them either. The number already sits on the key number of 7 at several books, and the Allen knee concern moved DraftKings and FanDuel down to 6.5 where New England is plus money at even price. Drake Maye practiced fully with a shoulder note. I see a mild lean on the points in a divisional game at 6.5 with even money, but nothing I would call a bet because the evidence beyond the price is thin and Allen's practice status is unresolved.

Total.Both sides have a case. Over: Allen and Buffalo score freely, and New England's defense has lost Barmore for practice and Gonzalez. Under: Patriots receivers are hurt, Allen's knee may limit scrambling, and Buffalo may lean on James Cook on the ground, which shortens the game. Weather is mild, about 66 degrees with light southwest wind, so it does not tilt anything. The number has come down from 49.5 and 50.5 openers to 48.5 everywhere, so the market has already priced the drop.